Cagrilintide vs Eloralintide

ON THIS PAGE

1. Key findings
2. The receptor family both act on
3. Where the two part company
4. The registry, and what its counts conceal
5. How much of each is actually published
6. What a buyer can check, and what nobody can
7. What this does not settle
8. Sources

Two amylin analogs, both lipidated peptides of roughly four and a half kilodaltons, both carrying registered phase 3 records, and one entered the trial registry almost two and a half years before the other. That gap decides how much of what can be read about either is a primary report rather than somebody restating a corporate announcement.

The molecular difference is the second thing, and it is sharper than the shared family name suggests. On the published in vitro record these two sit in different places on the same receptor family.

The receptor family both act on

Amylin receptors are not a single protein with a single gene. They are heteromers, the calcitonin receptor assembled with a receptor activity modifying protein, and the three combinations are named AMY1R, AMY2R and AMY3R. The 2021 pharmacology paper on cagrilintide states it plainly, describing amylin receptors as heteromeric assemblies combining calcitonin receptors with modifying proteins.

That architecture is why selectivity here is a real question rather than a marketing distinction. Every amylin receptor carries a calcitonin receptor inside it, so a compound built for one will reach the other unless its design stops it, and how much it reaches is a measurable property.

Where the two part company

Cagrilintide carried the code AM833 through its early characterization. The 2021 paper in the Journal of Pharmacology and Experimental Therapeutics, DOI 10.1124/jpet.121.000567, examined it across 25 pharmacological measures alongside six other agonists and calls it "a nonselective AMYR and CTR agonist" with "a unique pharmacological profile across diverse measures of receptor binding, activation, and regulation". A companion paper in the Journal of Medicinal Chemistry that August, DOI 10.1021/acs.jmedchem.1c00565, describes the molecule as a "stable, lipidated long-acting amylin analogue" and names the design problem behind it: native amylin aggregates into fibrils, and pramlintide, the one amylin analog ever approved in the United States, has a short half-life.

Eloralintide, code LY3841136, arrives at the same family from the other direction. Its originators at Eli Lilly report in Molecular Metabolism, December 2025, DOI 10.1016/j.molmet.2025.102271, that the compound "preferentially activated human AMY1R (12-fold > CTR, 11-fold > AMY3R)". The same paper carries the only direct comparison of the two molecules this search found in the published literature, and it is a rodent-lane finding: in lean rats, eloralintide "induced significantly less conditioned taste avoidance" than cagrilintide. One paper, one animal model, one endpoint. It is not a human comparison and nothing in the published record makes it one.

The registry, and what its counts conceal

An intervention search of clinicaltrials.gov on 2026-09-16 returned 44 studies for cagrilintide and 18 for eloralintide. Filtered to phase 3, the counts are 22 and 5. Novo Nordisk A/S is the lead sponsor on every cagrilintide record returned; Eli Lilly and Company on every eloralintide record. Neither compound has a single registered study sponsored by anybody outside its originator.

The 44 flatters cagrilintide. Adding a title filter for CagriSema to the same intervention search returns 28, meaning 28 of the 44 brief titles name the fixed combination product rather than the amylin analog alone. Anyone counting registered studies as a proxy for how much is known about cagrilintide by itself is counting the wrong set by a factor of about three.

The dates are where the age gap sits. NCT04940078 is the earliest record the cagrilintide search returns, first posted 2021-06-25. The earliest for eloralintide is NCT06143956, a master protocol first posted 2023-11-22. Twenty-nine months separate them. Both dates are floors, not origins: early work on either molecule may sit in the registry under a code number the search term no longer matches.

How much of each is actually published

PubMed answered 99 for cagrilintide and 9 for eloralintide on 2026-09-16. The 9 is the number worth sitting with. Of those nine records, three carry a clinical-trial publication type: a phase 1 report in Diabetes, Obesity and Metabolism, the Molecular Metabolism discovery paper, and a 48-week phase 2 report in The Lancet dated 6 December 2025. The other six are reviews, a network meta-analysis and commentary: other people reading those three.

No phase 3 report for eloralintide appears in that index. For cagrilintide the same index answers differently. The term narrowed by the publication type Clinical Trial, Phase III returns seven records, all of them inside the 99, and every one of the seven reports cagrilintide together with semaglutide, as the fixed combination or co-administered, rather than the amylin analog by itself. The earliest two are in the New England Journal of Medicine dated 14 August 2025, PMID 40544433 and PMID 40544432. The other five run from Hypertension in February 2026 through The Lancet and Lancet Diabetes and Endocrinology to August 2026, and their titles carry the trial names REDEFINE and REIMAGINE. One of the seven, the REIMAGINE 2 report, sets the combination against a cagrilintide arm; none of the seven is a report of the analog on its own.

So of the 27 registered phase 3 records, 22 for cagrilintide and 5 for eloralintide, the five eloralintide records have produced no journal report, while the cagrilintide programme's REDEFINE and REIMAGINE trials have been reported since August 2025. A phase 3 claim about eloralintide in September 2026 is sourced to a company statement or to nothing, because the index holds no phase 3 report for it. A phase 3 claim about cagrilintide has peer-reviewed reports behind it, and every one of those reports is a report of the combination.

What a buyer can check, and what nobody can

Identity is checkable on paper. Almost nothing else is. PubChem carries a computed record for each: cagrilintide at CID 171397054, molecular formula C194H312N54O59S2, average mass 4409 daltons, and eloralintide at CID 175663130, C201H319N49O65S2, 4526 daltons. Those two numbers are 117 daltons apart. A mass spectrometry line reporting observed against expected mass can therefore separate these two without difficulty, and a certificate carrying purity by HPLC alone cannot, because area percent describes how much of one peak eluted and says nothing about which peptide it was.

What cannot be checked is anything that would normally sit behind a compendial claim. A Drugs@FDA query for either name on 2026-09-16 returned no application record, so there is no approved article to act as a comparator for either. The class has exactly one approval to its name, pramlintide acetate under NDA 021332, originally approved 16 March 2005, and every product formulation listed under that application is marked Discontinued in the same database. So the reference point a laboratory would reach for does not exist as a marketed article either. A certificate for either compound is a statement by whoever wrote it, checkable against a computed mass and against nothing else.

What this does not settle

Registry counts measure registration, not knowledge: two of the 44 cagrilintide records carry a status of WITHDRAWN on the date read. PubMed counts measure indexing, so work in venues it does not index is invisible to the 99 and the 9 alike. The receptor figures quoted here are in vitro measurements made by the groups that developed each compound, which is the ordinary condition of early pharmacology and still worth naming. The PubChem masses are computed from deposited structures rather than measured on material, so they are the expected value an identity check compares against, not evidence about any vial. No supplier was examined for this piece and none is named.

The receptor difference is on the record and unlikely to move. What is still unreported is a narrower set than the 27 suggests: the five eloralintide phase 3 records, none of which has reached a journal, and the three cagrilintide phase 3 records whose brief titles name the analog without CagriSema, first posted on 24 October 2025, 24 October 2025 and 4 August 2026. Those are the lanes where a phase 3 claim in September 2026 has no journal report behind it at all.

Sources

Everything above was read on 2026-09-16. Study counts, sponsors, phases, NCT numbers and first-posted dates: the clinicaltrials.gov public API, one intervention search per compound name, the phase 3 figures from those searches filtered to phase 3, the CagriSema figure from the cagrilintide search with a title filter added. Literature counts and publication types: PubMed, term searches for each name, and the same terms narrowed by the publication type Clinical Trial, Phase III, which returned seven records for cagrilintide and none for eloralintide. The three cagrilintide phase 3 registry records that do not name the combination: the phase 3 search above with brief titles matched case-insensitively against CagriSema, which leaves three of the 22. Receptor characterization: Journal of Pharmacology and Experimental Therapeutics, June 2021, DOI 10.1124/jpet.121.000567; Journal of Medicinal Chemistry, 12 August 2021, DOI 10.1021/acs.jmedchem.1c00565; Molecular Metabolism, December 2025, DOI 10.1016/j.molmet.2025.102271. The phase 2 report: The Lancet, 6 December 2025, PMID 41207310. Formulas and computed masses: PubChem, CID 171397054 and CID 175663130. Application status: the Food and Drug Administration's Drugs@FDA database through the agency's public openFDA endpoint, which returned no record for either investigational compound and one, NDA 021332, for pramlintide acetate.